[OpenAFS] OpenAFS and Raid
Hartmut Reuter
reuter@rzg.mpg.de
Tue, 18 Dec 2001 17:43:03 +0100
I do not understand why the availability of all user homedirs should
depend on a single machine. The volume where the mountpoints for the
user volumes are in is replicated, of course. The "backup-machine" with
the additional RO-volumes is an add-on which increases availability or,
at least, reduces possible down-times. Our cell is not really small (17
TB)!
Hartmut
Derek Atkins schrieb:
>
> Yes, but I suspect that's not the behavior you would want.
> Unless you have a relatively small cell, do you REALLy want
> all of your user homedir availability to be dependent upon
> a single machine?
>
> -derek
>
> "Neulinger, Nathan" <nneul@umr.edu> writes:
>
> > You could just mount /afs/mpg.de/user RW. That would force it.
> >
> > -- Nathan
> >
> > ------------------------------------------------------------
> > Nathan Neulinger EMail: nneul@umr.edu
> > University of Missouri - Rolla Phone: (573) 341-4841
> > Computing Services Fax: (573) 341-4216
> >
> >
> > > -----Original Message-----
> > > From: Derek Atkins [mailto:warlord@MIT.EDU]
> > > Sent: Tuesday, December 18, 2001 9:40 AM
> > > To: Hartmut Reuter
> > > Cc: Petros Triantafyllidis; OpenAFS Info List
> > > Subject: Re: [OpenAFS] OpenAFS and Raid
> > >
> > >
> > > Hartmut,
> > >
> > > How do you prevent the /afs/mpg.de/user/foo from accessing the
> > > readonly mountpoint? Do you specifically mount all user.foo
> > > volumes -rw?
> > >
> > > -derek
> > >
> > > Hartmut Reuter <reuter@rzg.mpg.de> writes:
> > >
> > > > Petros Triantafyllidis schrieb:
> > > > >
> > > > > Hi all,
> > > > > We are planning to install OpenAFS 1.2.2a on a new
> > > server running RH
> > > > > 7.1, kernel 2.4.16smp. After the testing period this
> > > machine will be
> > > > > the second server of a primary one which runs Solaris 2.6
> > > and Transarc
> > > > > AFS 3.6.
> > > > >
> > > > > The thought is to implement software RAID on the Linux
> > > box. The question
> > > > > is whether raid-5 can co-live with AFS. What are the
> > > possible pitfalls and
> > > > > how should it be done in order we obtain the maximum of
> > > performance
> > > > > (assuming redundancy is a must)?
> > > >
> > > > We are running a machine with SuSE 7.3 (kernel 2.4.10) with
> > > 2 RAID5s,
> > > > LVM and a huge reiserfs with 900 GB on top of this. Reading
> > > and writing
> > > > performance seems to be ok (both ~ 36 MB/s). The only
> > > problem I have is
> > > > that the RAID seems not to detect a bad disk by itself.
> > > When a disk dies
> > > > we get a lot of SCSI-timeout messages in /var/log/messages, but it
> > > > doesn't understand that the disk is dead. Worse: we have to
> > > reboot after
> > > > a "raidsetfaulty" for the disk in order to get him to
> > > rebuild the RAID.
> > > > We use this partition for readonly-volumes of the user's
> > > > home-directories. I have written a "vos convert" command (MR-AFS) to
> > > > convert a RO-volume to a RW-volume. This allows to come up after the
> > > > loss of a partitioon much faster than if you had to restore
> > > the volumes
> > > > from tape. So we do each night a "vos release" instead of a "vos
> > > > backupsys".
> > > >
> > > > Hartmut
> > > >
> > > >
> > > > >
> > > > > Thanks a lot in advance,
> > > > > Petros
> > > > >
> > > > > --
> > > > > Dr. TRIANTAFYLLIDIS PETROS E-MAIL: trian@itc.auth.gr
> > > > > ^^^^^^^^^^^^^^^^^^^^^^^^^^ http://afs.itc.auth.gr/~trian
> > > > > AFS Technical Support Team
> > > > > Aristotle University - Information Technology Center (KYTP)
> > > > > POBox 888,54006 Thessaloniki-GREECE-TEL:+30-31-998444,FAX:998302
> > > > >
> > > > > _______________________________________________
> > > > > OpenAFS-info mailing list
> > > > > OpenAFS-info@openafs.org
> > > > > https://lists.openafs.org/mailman/listinfo/openafs-info
> > > >
> > > > --
> > > > -----------------------------------------------------------------
> > > > Hartmut Reuter e-mail reuter@rzg.mpg.de
> > > > phone +49-89-3299-1328
> > > > RZG (Rechenzentrum Garching) fax +49-89-3299-1301
> > > > Computing Center of the Max-Planck-Gesellschaft (MPG) and the
> > > > Institut fuer Plasmaphysik (IPP)
> > > > -----------------------------------------------------------------
> > > > _______________________________________________
> > > > OpenAFS-info mailing list
> > > > OpenAFS-info@openafs.org
> > > > https://lists.openafs.org/mailman/listinfo/openafs-info
> > >
> > > --
> > > Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
> > > Member, MIT Student Information Processing Board (SIPB)
> > > URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH
> > > warlord@MIT.EDU PGP key available
> > > _______________________________________________
> > > OpenAFS-info mailing list
> > > OpenAFS-info@openafs.org
> > > https://lists.openafs.org/mailman/listinfo/openafs-info
> > >
> > _______________________________________________
> > OpenAFS-info mailing list
> > OpenAFS-info@openafs.org
> > https://lists.openafs.org/mailman/listinfo/openafs-info
>
> --
> Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
> Member, MIT Student Information Processing Board (SIPB)
> URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH
> warlord@MIT.EDU PGP key available
> _______________________________________________
> OpenAFS-info mailing list
> OpenAFS-info@openafs.org
> https://lists.openafs.org/mailman/listinfo/openafs-info
--
-----------------------------------------------------------------
Hartmut Reuter e-mail reuter@rzg.mpg.de
phone +49-89-3299-1328
RZG (Rechenzentrum Garching) fax +49-89-3299-1301
Computing Center of the Max-Planck-Gesellschaft (MPG) and the
Institut fuer Plasmaphysik (IPP)
-----------------------------------------------------------------